Ứng dụng các kỹ thuật khai phá dữ liệu trên các cở sở dữ liệu liên quan đến giấy phép lái xe

Phóng nhanh Gãy chân Phóng nhanh Gãy chân [22,35] Phóng nhanh Gãy tay... Dim Mang as TANSUATDim i,j , f As Integer, sName As string Dim sSel as string //... Private Sub Node_LinkByVal I

M C L C

L .i

TRANG THÔNG TIN LU ii

M C L C iv

DANH M C CÁC T VI T T T vi

DANH M C B NG BI U vii

DANH M C HÌNH viii

M U 1

1 Lý do ch tài 1

2 M c tiêu và nhi m v nghiên c u 2

ng và ph m vi nghiên c u 2

u 2

5 Ý c và th c ti n c tài 3

6 B c c lu 3

T NG QUAN V KHAI PHÁ D LI U 4

4

1.2 Các d ng d li u có th khai phá 4

ng ti p c n và k thu t áp d ng 6

c ng d ng 7

1.5 M t s li u 7

n và quy n p 7

1.5.2 Cây quy nh và lu t: 8

n 8

1.5.4 M ng neural 8

1.5.5 Gi i thu t di truy n 9

LU T K T H P VÀ THU T TOÁN TÌM LU T LU T K T H P 10

2.1 T ng quan v lu t k t h p 10

2.2 M t s khái ni n 11

2.2.1 Lu t k t h p 11

h tr c a m t t p m c (itemset) 11

2.2.3 T p m c ph bi n (frequent itemset): 11

a m t lu t: 11

2.3 Bài toán khai phá d li u b ng lu t k t h p và các pha th c hi n 12

2.4 M t s tính ch t c a t p m c ph bi n và lu t k t h p 14

2.4.1 M t s tính ch t v i t p m c ph bi n 14

2.4.2 M t s tính ch t c a lu t k t h p 14

2.5 Các lo i lu t k t h p 15

2.6 Các thu t toán khai phá d li u nh lu t k t h p 17

2.6.1 Khai phá lu t k t h u t d li u tác v 17

2.6.2 Khai phá lu t k t h ng 37

NG D NG LU T K T H P KHAI PHÁ D LI U TAI N N GIAO THÔNG 40

3.1 M u c a bài toán 40

3.2 K t qu c và phát tri n ng d ng 58

K T LU N 59 TÀI LI U THAM KH O

TT

TTTT

DANH M C B NG BI U

B ng 2.2 tin c y c a m t s lu t k t h c sinh ra t b ng 2.1 14

B ng 2.5 Minh h a k t qu c 1 c a thu t toán EP-growth 30

M U

1

v.v

mình

g tin

2008

3.1

S

2

3

Ch ng 1

T NG QUAN V KHAI PHÁ D LI U

1980 xem nh

-processing and Preparation)

xem nh thông tin d

Ph

ng

1.5

là các u

nh

1.5

cong ng

1.5

T t

Ch ng 2

LU T K T H P VÀ THU T TOÁN TÌM LU T LU T K T H P 2

Y Z => X Y X Z

Y và Y Z Z

gãy tay" (L1)

l

(L2)

(L3)máy

ng

(L4)

(L5)

t

Ck = sinh_ung_cu(Lk-1); //by joining Lk-1 with itself For

2 Lk-1 và l2 <> l1Begin

X

{B}, {D}

sau: B conf = 100%

forall large itemsets lk, k 2 do begin

if(conf minconf) then

output the rule (lk-hm+1) hm+1

genrules (ABCDE, ACDE)

BE, ADE BC, CDE BA, ACE BD

BE và ABCD

2

Tuy nhiên trong tr

y nh sau:

1N2 Nm

Cây T là FP-tree

Var i,j,k: Integer;

Var varNode: Node;

-path =-path & varNode.Name varnode=T.Nodes(varnode.ParentID)

FP ngay

2

phân

ID Nguyên nhân

Phóng nhanh Gãy chân

Phóng nhanh Gãy chân

[22,35] Phóng nhanh Gãy tay

TNGT

TNGT

Các 3.1

Dim Mang() as TANSUAT

Dim i,j , f As Integer, sName As string

Dim sSel as string

//

Redim Mang(rsMining.Fields.Count -1)

Do While Not rsMining.EOF

For i =0 to rsMining.Fields.Count -1

If Not isnull(rsMining.Fields(i).Value) then

Mang(i).Name =rsMining.Fields(i).Name Mang(i).Tansuat

L For i=0 to Ubound(Mang(i))

If Mang(i).Tansuat < minsupp Then

Redim Preserve

Mang(i-End if

sSel=sSel & Mang(i).Name

Next

-Public Type myTree

'Out put: Cây

FP-Dim m As Long, n As Long, k As Long

Dim strTemp As String

m = UBound(myTree(iRoot).Con)

If m > 0 Then

For i = 1 To m

k = myTree(iRoot).Con(i)

If myTree(k).Ten = Mid(strItem, 1, 3) Then

myTree(k).Tansuat = myTree(k).Tansuat + myifrequent

ReDim Preserve myTree(iRoot).Con(m + 1)

ReDim Preserve myTree(k).Con(0)

ReDim Preserve myTree(iRoot).Con(m + 1)

Private Sub Node_Link(ByVal Item As String, ByVal iPos As Long, ByRefMang() As Tansuat)

Dim i As Integer, j As Integer

A>{} hay {}

-Private Sub FP_Growth(ByRef HeaderTable() As Tansuat, myT() As myTree, ByVal Item As String)

Dim i As Long, j As Long, k As Long

Dim m As Long, n As Long, h As Long

Dim myString As String, strTemp As String

Dim fixId As Long, ifrequent As Long

Dim myHead() As Tansuat

Dim myCay() As myTree

Dim strInsert As String, strValue As String

Dim bData As Boolean

Call Generate_Itemset(strTemp, Item, myItemset, myT)

'Lay ra chuoi chua ta cac ten cac nut tren cay

'Tao mang chua tat ca cac tap con co the cua tap cac nut noi tren voi do ho tro la min trong tat ca

'Ket qua cuoi cung la duoc mang chua cac tap muc pho bien 'Bo qua

luat A->{} va nguoc lai

Else

For i = 0 To n

bData = False

k = UBound(HeaderTable(n - i).Vitri)

cn.Execute "Delete from Mytree"

'Duyet cay day vao CSDL

For j = 1 To k

strInsert = strInsert & "," & myT(h).Ten

strValue = strValue & ",1"

strInsert = "Insert into MYTREE(mau,Tansuat," &strInsert & ")"

strValue = " Values('" & strTemp & "'," & ifrequent &

"," & strValue & ")"

cn.Execute strInsert & strValue

Call Create_List("MyTree", myHead)

'Duyet CSDL lan 2 tao Cay phu thuoc

If UBound(myHead) = 0 And myHead(0).Tansuat < minsupp Then 'Khong lam

gi ca nua

Else

'Neu cay phu thuoc khong rong thi goi de quy

myString = myString & strTemp

ReDim myCay(0)

Call Tao_cay("MyTRee", myCay, myHead)

If UBound(myCay) > 0 Then Call FP_Growth(myHead, myCay, myString)End If

Set rsTemp = New ADODB.Recordset

rsTemp.Open "Select * from " & strTableName, cn, adOpenForwardOnly, adLockReadOnly

If rsTemp.RecordCount = 0 Then Exit Sub rsTemp.MoveFirst

ReDim myArray(rsTemp.Fields.count - 3) Do While Not rsTemp.EOF

'S¾p xÕp

For i = 0 To UBound(myArray)

For j = i + 1 To UBound(myArray)

If myArray(j).Tansuat > myArray(i).Tansuat Then

f = myArray(i).Tansuat strName = myArray(i).Name myArray(i).Name =

myArray(j).Name myArray(i).Tansuat = myArray(j).Tansuat myArray(j).Name = strName myArray(j).Tansuat = f

End If

Next

Next

For i = 0 To UBound(myArray)

ReDim Preserve myArray(i).Vitri(0)

If myArray(i).Tansuat < minsupp Then

If UBound(myT(i).Con) > 1 Then SingleTree = False

Exit Function End If

'MsgBox strItem & ":Mau:" & Mau

Dim strTemp As String, myItem As String Dim i As Integer, j As Integer, k As Integer

Dim m As Integer, n As Integer, h As Integer, l As Integer Dim myArray() As MySet

Dim A As Long, B As Long, x As Long, y As Long, count As Long

m = Len(strItem) / 3 ' Sè phÇn tö cña m¶ng mySet Mçi Item gåm 3 ký tù

If m = 0 Then Exit Sub

ReDim Preserve myArray(m)

'ReDim myArray(0).Item(0)

'myArray(0).Item(0) = Mau

ReDim Preserve myArray(1).Item(m)

ReDim Preserve myArray(1).Tansuat(m)

For i = 1 To m

myArray(1).Item(i) = Mid(strItem, i * 3 - 2, 3)

Next

For i = 2 To m

ReDim Preserve myArray(i).Item(0)

ReDim Preserve myArray(i).Tansuat(0)

n = 0: k = 0

n = UBound(myArray(i - 1).Item)

For j = 1 To n

myItem = Right(myArray(i - 1).Item(j), 3)

k = InStr(1, strItem, myItem, vbTextCompare) + 3 strTemp

= ""

strTemp = Mid(strItem, k)

l = Len(strTemp) / 3

If l > 0 Then

For h = 1 To l

ReDim Preserve myArray(i).Item(UBound(myArray(i).Item) + 1)

ReDim Preserve myArray(i).Tansuat(UBound(myArray(i).Tansuat) + 1)myArray(i).Item(UBound(myArray(i).Item)) =\myArray(i -

1).Item(j) & Mid(strTemp, (h - 1) * 3 + 1, 3)I Next

ReDim Preserve myMang(k)

ReDim myMang(k).Item(UBound(myArray(i).Item)) ReDim

Dim rsLan2 As ADODB.Recordset

Dim i As Long, m As Long, j As Long, k As Long

Dim arrTemp() As String

Dim myString As String, strSel As String

'Nót gèc cã nh·n rçng, cha=0, Kh«ng cã con ghÜa lµ T(0).Con(0)=0, NÕu cã con th× T(i).con(j)>0 víi j>=1

If Not IsNull(.Fields(i).Value) Then

K T LU N

trình

:

Khoa CNTT, B môn HTTT Kinh T , Thái Nguyên

Ti ng Anh:

[9] Alan Rea (1995), Data Mining An Introduction, The Parallel Computer Center,

[10] Jiawei Han and Micheline Kamber (2002), Data mining: Concepts and Techniques, University of Illinois, Morgan Kaufamn Publishers

[11] Rakesh Agrawal, Tomasz Imielinski and Arun Swami (1993), Mining association rules between sets of Items in lagre databases, In Proc of the ACM SIGMOD Conference on Mangement of Data, Washington, DC

[12] Mohammed J.Zaki and Ching Jui Hsiao (1999), CHAM: An Efficient Algorithm for Closed Association Rules Mining, PRI Technical Report 99 [13] Jiawei Han and Micheline Kamber, Data mining: Concepts and Techniques, Academic Press 2001

[14] Zaki M., Parthasarathy S., Ogihara M.: New algorithms for fast discovery of association rule, KDDM 1998

[15] www.mimuw.edu.pl/~son/datamining